Related Experiment Video
Updated: Aug 26, 2026
A Standardized Obstacle Course for Assessment of Visual Function in Ultra Low Vision and Artificial Vision
Published on: February 11, 2014
A Within-Population Comparison of the Psychometric Properties and Clinical Feasibility of Visual Function
Micaela Gobeille1, Nicole Ross1,2, Chris Bradley3
1New England College of Optometry, Department of Specialty Care and Vision Science, Boston, MA, USA.
Purpose:
The purpose of this study was to determine which of three visual function questionnaires (VFQs) is most appropriate and practical for low vision (LV) clinical use when administered to the same LV patient sample.
Methods:
Activity Inventory goals (AI-G), Veterans Affairs Low-Vision Visual Functioning Questionnaire 48 (VA LV VFQ-48), and National Eye Institute Visual Function Questionnaire 25 visual function subscale (NEI VFQ-25c-VF) were administered to 74 patients with LV pre- and 6 months post-LV examination. Rasch analysis estimated person measures, item measures, and infit mean square statistics (infits). Targeting was assessed using the Kolmogorov-Smirnov (KS) test, while Deming regression and Pearson correlation tests compared person measures between VFQs and between pre- and post-LV examination. The Wilcoxon signed-rank test compared administration time per item for a subset (n = 30) of participants.
Results:
Statistically significant ceiling effects were observed for AI-G and VA LV VFQ-48, but not for NEI VFQ-25c-VF (P = 0.39). However, NEI VFQ-25c-VF had the highest percentage of combined pre- and post-LV person (31.1%) and item (20%) infits beyond a desirable range of 0.5 to 1.5, indicating greater multidimensionality. Person measures were moderately correlated between AI-G and both VA LV VFQ-48 (r = 0.64) and NEI VFQ-25c-VF (r = 0.63), but strongly correlated between NEI VFQ-25c-VF and VA LV VFQ-48 (r = 0.85). Regarding clinical practicality, pre-LV examination administration time per item was significantly faster for VA LV VFQ-48 than for AI-G (P = 0.00016) or NEI VFQ-25c-VF (P = 0.0065).
Conclusions:
No VFQ was superior in all respects. Although NEI VFQ-25c-VF had the best targeting, its higher multidimensionality supports using VA LV VFQ-48 if patient ability remains below the instrument ceiling.
Translational Relevance:
Improvements to widely used VFQs are required for use in LV clinics.
